Complex product development time prediction and optimization method based on design structure matrix

Abstract

The invention discloses a complex product development time prediction and optimization method based on a design structure matrix. The complex product development time prediction and optimization method is characterized in that for a complex product development process composed of n design activities with known average completion time and known information dependence relationship, all design activities of the complex product development process form a design activity set; a design structure matrix corresponding to a product development project is constructed, and an optimization objective function is constructed: according to the optimization objective function, the design structure matrix corresponding to the product development project is solved. The complex product development time prediction and optimization method takes a complex product development project as a research object, aims to predict and reduce product development time, performs analysis and modeling on the DSM corresponding to the product development project, identifies the first-order rework and the second-order rework in the product development project, obtains the optimal execution sequence of all design activities involved in the first-order rework and the second-order rework through optimization, and finally achieves the minimum total rework time of the product development project, so that the purposes of reducing the total product development time, reducing the total product development cost and improving the product development success rate are achieved.

Description

technical field [0001] The invention relates to a complex product development time prediction and optimization method based on a design structure matrix. Background technique [0002] As the market competition environment becomes increasingly fierce, the product life cycle continues to shorten, and the replacement becomes faster and faster. Improving the efficiency of new product development has become an important means for modern enterprises to maintain their core competitiveness in the market and achieve commercial success. Research data show that: for most companies, more than 50% of their profits come from the latest products; but on the other hand, about 70% of product development projects exceed the estimated time schedule, and the planned delivery time is overdue by an average of 20 % to 50%, more than 90% of product development projects spend over budget, and only about 60% of product development projects are ultimately successful. This is because the actual produc...

AI Technical Summary

Problems solved by technology

[0007] However, these objective functions also have some flaws that prevent them from being fully effective in the analysis of rework in product development projects: (1) These proxy objective functions do not establish a quantifiable relationship with the final goal (reducing the total time of product development projects). In other words, it is difficult to find the corresponding management significance of the final value of the proxy objective function in the actual project, and it is difficult for the project manager to estimate the execution time and optimization degree of the product development project based on the optimization result of the proxy objective function. Reasonably allocate personnel and resources and arrange rush work

(2) The existing proxy objective functions do not fully explore all the information in the DSM matrix. Almost all proxy objective functions use the feedback information in the DSM matrix (above the diagonal), and for the feedforward information (diagonal Below the line) there is little research, but feed-forward information is also an important reason for rework iterations, especially high-order rework, which inevitably leads to the lack of convincing accuracy of existing algorithms

(3) Existing surrogate objective functions have insufficient research on high-order rework

Although the model covers all first-order rework and higher-order rework in the DSM matrix, it has high accuracy, but the complexity of the model is not convenient for optimizing large-scale product development projects in practice

Banerjee et al. designed a SEPTR heuristic algorithm based on this model, but this algorithm still cannot solve the large-scale activity sequencing problem, and it takes an average of 37 minutes to solve the DSM matrix of 9 design activities

[0009] To sum up, the existing optimization algorithms are insufficient to mine a large number of complex information dependencies among design activities in product development projects, and do not consider high-level rework in design iterations, so it is difficult to ensure accurate and effective optimization results

In addition, the optimized values ​​obtained by these algorithms have no practical management significance, and it is difficult to analyze and utilize these optimized values

Images